Core and advanced AI

The maths and research behind modern AI

Course 10 of 10, Adults

1. Vectors, matrices and tensors: The language of neural networks
2. Loss and gradient descent: How models get less wrong
3. Backpropagation: Sharing out the blame
4. The transformer, layer by layer: The architecture behind modern LLMs
5. Scaling laws: Why bigger models got better
6. RLHF and preference tuning: Teaching models what people prefer
7. Reinforcement learning: Agents, rewards and policies
8. How diffusion models work: From noise to images
9. Interpretability: Looking inside the black box
10. Alignment and AI safety: Making AI do what we really intend